SpEDIT: A fast and efficient CRISPR/Cas9 method for fission yeast

The CRISPR/Cas9 system allows scarless, marker-free genome editing. Current CRISPR/Cas9 systems for the fission yeast Schizosaccharomyces pombe rely on tedious and time-consuming cloning procedures to introduce a specific sgRNA target sequence into a Cas9-expressing plasmid. In addition, Cas9 endonuclease has been reported to be toxic to fission yeast when constitutively overexpressed from the strong adh1 promoter. To overcome these problems we have developed an improved system, SpEDIT, that uses a synthesised Cas9 sequence codon-optimised for S. pombe expressed from the medium strength adh15 promoter. The SpEDIT system exhibits a flexible modular design where the sgRNA is fused to the 3’ end of the self-cleaving hepatitis delta virus (HDV) ribozyme, allowing expression of the sgRNA cassette to be driven by RNA polymerase III from a tRNA gene sequence. Lastly, the inclusion of sites for the BsaI type IIS restriction enzyme flanking a GFP placeholder enables one-step Golden Gate mediated replacement of GFP with synthesized sgRNAs for expression. The SpEDIT system allowed a 100% mutagenesis efficiency to be achieved when generating targeted point mutants in the ade6 + or ura4 + genes by transformation of cells from asynchronous cultures. SpEDIT also permitted insertion, tagging and deletion events to be obtained with minimal effort. Simultaneous editing of two independent non-homologous loci was also readily achieved. Importantly the SpEDIT system displayed reduced toxicity compared to currently available S. pombe editing systems. Thus, SpEDIT provides an effective and user-friendly CRISPR/Cas9 procedure that significantly improves the genome editing toolbox for fission yeast.


Introduction
The fission yeast Schizosaccharomyces pombe is a powerful model organism widely used in cellular and molecular biology (Fantes & Hoffman, 2016). Traditionally, gene manipulation in S. pombe is achieved by transforming a DNA construct that includes the desired change alongside a selectable marker. The DNA construct integrates in the genome via flanking regions that target the genomic locus of interest. DNA constructs vary depending on the application but commonly consist of a sole PCR product that comprises an insertion, deletion or tagging cassette amplified from an existing plasmid (Bähler et al., 1998). Albeit convenient, this approach results in a selectable marker integrated at the target locus, consequently disrupting the local genomic context and limiting the availability of markers for subsequent manipulations.
The prokaryotic CRISPR/Cas9 system enables flexible and scarless genome editing without the necessity of selectable markers escorting the introduced DNA change and disturbing the local genomic environment (Hsu et al., 2014;Jinek et al., 2012). Adapted from a genome defence mechanism against invading DNA, the engineered minimal CRISPR/Cas9 system consists of a Cas9 endonuclease and a single-guide RNA (sgRNA) chimera that contains both the trans-activating CRISPR RNA and the targeting CRISPR RNA (Jinek et al., 2012). The sgRNA sequence targets the system to a defined genomic location where the Cas9 endonuclease binds to a protospacer adjacent motif (PAM). The Cas9 enzyme then creates a double-strand break (DSB) three base pairs upstream of the PAM site in the protospacer sequence (Jinek et al., 2012). Following DSB generation, repair is executed either through error-prone non-homologous end-joining (NHEJ), where strand end resection and subsequent repair frequently induce indels, or through high-fidelity homology-directed repair (HDR). HDR involves recombination via sequence homology and therefore can be exploited to generate precise mutations by providing a DNA editing template that contains the required DNA change and engages in homologous recombination (HR) with the cleaved region (Hsu et al., 2014).
Implementation of the CRISPR/Cas9 system in S. pombe has been previously described in several reports ( To date, all CRISPR/Cas9 systems for S. pombe utilise the promoter/leader sequence of K RNA (rrk1) and a hammerhead ribozyme (HHR) for sgRNA cassette expression (Jacobs et al., 2014). A humanised Cas9 endonuclease expressed from the strong adh1 promoter was used in the original S. pombe system (Jacobs et al., 2014). The resulting high levels of the Cas9 endonuclease were found to be detrimental for S. pombe growth. This Cas9 toxicity was partially bypassed by co-expression of the sgRNA and the Cas9 enzyme from a single plasmid (Jacobs et al., 2014).
Cloning of a sgRNA target into the single sgRNA/Cas9 plasmid was originally executed via CspCI digestion (Jacobs et al., 2014). This procedure proved to be extremely inefficient due to the large plasmid size and the inconsistency in available commercial CspCI preparations (Rodríguez-López et al., 2016). A subsequent study attempted to overcome this problem by implementing a PCR-based method in which a sgRNA target is introduced into a single sgRNA/Cas9 plasmid by using overlapping PCR primers that contain the sgRNA sequence (pMZ379 plasmid, Rodríguez-López et al., 2016). Although an improvement over the initial system, this PCR-based method generated only a low frequency of bacterial colonies that contain correct and intact constructs. As a consequence, the required screening makes the entire process inefficient and time consuming.
To circumvent issues pertaining to Cas9 toxicity and inefficient sgRNA cloning procedures, here we report the development of SpEDIT, an improved CRISPR/Cas9 system for the efficient manipulation of the fission yeast genome. SpEDIT employs a highly effective one-step Golden Gate cloning strategy for the insertion of sgRNAs, that in combination with the use of a GFP placeholder allows visual screening for identification of positive clones. The Cas9 endonuclease gene implemented in this system is codon-optimised for expression in S. pombe and driven by the medium strength promoter adh15, resulting in reduced toxicity associated with Cas9 levels. SpEDIT can generate targeted ade6 and ura4 point mutants in asynchronous cells with 100% mutagenesis efficiency. Moreover, SpEDIT allows simultaneous editing at two non-homologous genes at distinct locations in the S. pombe genome, as well as seamless insertion, deletion and tagging at S. pombe loci. SpEDIT provides an efficient and simple CRISPR/Cas9 method to easily manipulate the genome of the fission yeast.

Results
The SpEDIT system The SpEDIT system has been developed to address the two main complications associated with existing CRISPR/Cas9 methods for S. pombe: toxicity associated with Cas9 overexpression, and laborious cloning procedures required to insert a specific sgRNA target sequence into a Cas9-containing plasmid. An overview of the SpEDIT system is provided (Figure 1) along with a full protocol (see Methods).
High levels of human codon-optimised Cas9 endonuclease constitutively expressed from the exceptionally strong adh1 promoter (400 RNA molecules/cell, PomBase, Lock et al., 2019) lead to reduced cell growth in S. pombe (Jacobs et al., 2014). A recent report attempted to solve this toxicity problem by expressing the human codon-optimised Cas9 under control of the repressible nmt41 promoter (Hayashi & Tanaka, 2019). Although this approach does generate mutations, it requires the non-ideal use of minimal media and relies on auxotrophic (ura4or leu1 -) strains to allow plasmid selection. Moreover, the mutagenesis efficiency obtained was dependent on the selectable marker employed.
To overcome the toxicity related to high levels of humanised Cas9, we synthesized a Cas9 gene codon-optimised for expression in S. pombe (SpCas9) that is transcribed from the medium strength adh15 promoter (Yamagishi et al., 2008). This adh15-SpCas9 gene is carried on a new plasmid, pLSB, that contains a choice of dominant selectable markers. Versions bearing natMX6, kanMX6 or hphMX6 markers are available, thereby allowing the SpEDIT system to be employed on fission yeast strains that harbour various manipulations where other selectable markers are already present (Figure 2A).
Eukaryotic CRISPR/Cas9 systems usually rely on snRNA or snoRNA RNA polymerase III (RNAPIII) promoters for transcription of the sgRNA cassette (Cong et al., 2013). However, characterised S. pombe RNAPIII genes contain promoter elements within the transcribed region, preventing their use for generating accurately positioned 5' ends, and RNA polymerase II (RNAPII) promoters generally do not generate transcripts with precise 5' and 3' ends. Consequently, all sgRNA expression systems for S. pombe so far utilise the rrk1 promoter plus its downstream 5' untranslated region which generates a RNAPII transcript with a cleavable leader RNA (Jacobs et al., 2014). Insertion of sgRNA sequences targeting genomic regions of interest into the rrk1 sgRNA expression cassette in current CRISPR/Cas9 systems for S. pombe relies on slow and arduous cloning procedures involving either traditional restriction digestion (Hayashi & Tanaka, 2019; Jacobs et al., 2014) or PCR over a long template (Rodríguez-López et al., 2016). An alternative method that uses in vivo gap repair to assemble a gapped Cas9-encoding plasmid and a PCR-amplified sgRNA fragment into a single circular plasmid has been reported (Zhang et al., 2018). Although this method provided an advance, this system still utilises humanised Cas9 expressed from the very strong adh1 promoter, which consequently reduces cell growth due to Cas9-associated toxicity.
It has previously been shown that the upstream tRNA Ser gene of an S. pombe tRNA Ser -tRNA Met gene pair drives RNAPIII expression of the downstream tRNA Met gene (Hottinger-Werlen et al., 1985). We therefore use this tDNA Ser to drive sgRNA expression in S. pombe. The resulting SpEDIT system employs a modular design where sgRNAs are expressed from this tDNA Ser sequence and fused to the hepatitis delta virus ribozyme (HDV), as previously described for Saccharomyces cerevisiae (Ryan et al., 2014). The tDNA Ser acts as a RNAPIII promoter and the self-cleaving ribozyme protects and defines the 5' end of the resulting sgRNA (Figure 2A and B). The presence of the HDV ribozyme was shown to result in a six-fold increase in sgRNA abundance and this correlated with high targeting efficiency in S. cerevisiae (Ryan et al., 2014). To facilitate cloning of sgRNA target sequences into this tDNA/HDV expression cassette, we placed sites for the BsaI type IIS restriction enzyme on each side of a GFP placeholder, thereby allowing one-step insertion of sgRNAs via Golden Gate cloning. Importantly, this strategy also permits visual screening to identify colonies that have lost the green GFP fluorescence signal indicating that the GFP has been successfully replaced with an incoming sgRNA ( Figure 2B and C).
SpEDIT can generate targeted ade6 and ura4 point mutants in asynchronous cells with 100% mutagenesis efficiency To assess the performance of the SpEDIT system in comparison to the existing pMZ379 system (Rodríguez-López et al., 2016), we targeted the ade6 + and ura4 + genes and provided HR templates that disable the PAM (NGG) sequence downstream of the sgRNA target to generate premature STOP codon mutations ( Figure 3A). ade6 + and ura4 + mutations can be easily scored due to their characteristic phenotypes: ade6 mutants, pink colonies develop on low (1/10) adenine-containing plates; ura4 mutants, cannot grow in the absence of supplementing uracil (uracil auxotrophy) but can grow in the presence of counterselective 5-fluoroorotic acid (FOA resistant) ( Figure 3B).
We scored cloNAT-resistant colonies after electroporation of asynchronous cultures with either SpEDIT/pLSB or pMZ379 plasmids expressing sgRNA designed to mediate cleavage within the ade6 + or ura4 + genes in the presence or absence of an HR template homologous to ade6 + or ura4 + , respectively. The results revealed that both pLSB and pMZ379 plasmids can generate targeted mutations in ade6 + and ura4 + with 100% efficiency when a matching HR template was co-transformed ( Figure 3C). However, when an HR template targeting a different gene or no HR template was provided, the number of cloNAT-resistant colonies obtained was dramatically reduced. This decrease in transformant number in the absence of an HR template is consistent with futile cleavage-repair cycles where the persistence of a double strand break prevents cell division and thus colony formation ( Figure 3C).
Sequence analysis of the resulting ade6 and ura4 mutants showed that when a matching HR template was co-transformed, all clones analysed harboured the mutation provided by the  Figure 1 or at allshirelab.com/spedit. Versions with natMX6 (cloNAT), kanMX6 (G418) or hphMX6 (hygromycin) S. pombe resistance markers are available. A Cas9 codon optimised for S. pombe (SpCas9) is expressed from the adh15 promoter (Padh15). B. Diagram of sgRNA cassette and cloning procedure. sgRNA cassette expression is driven by a tRNA Ser Pol III promoter (purple block arrow). A self-cleaving hepatitis delta virus (HDV) ribozyme is located at the 5' end of the sgRNA cassette (Ryan et al., 2014) (red block arrow). A superfolder green fluorescent protein (sfGFP) is used as placeholder (green block arrow). BsaI sites flanking sfGFP allow one-step insertion of a sgRNA target (light blue block arrow) into the sgRNA scaffold (grey block arrow) via Golden Gate cloning. The Pol III terminator sequence from S. cerevisiae SUP4 (tRNA Tyr ) is present at the 3' end of the sgRNA cassette (black block). C. The sfGFP placeholder allows cultures carrying empty (green) pLSB plasmids to be distinguished from sgRNA-loaded (non-green) pLSB plasmids. sgRNA, single guide RNA. A. Schematic of experiment to generate targeted ade6 and ura4 point mutants. A sgRNA-loaded pLSB or pMZ379 plasmid was co-transformed with an HR template that creates a premature STOP codon by disabling the PAM (NGG) sequence. sgRNA and HR template sequences for ade6 and ura4 are shown. Full HR template sequences can be found in Table 1. B. After transformation, cloNAT-resistant colonies were picked and re-streaked to non-selective YES plates. Cells were then replica-plated to indicated media to assess their phenotype. Representative plates from two independent experiments are shown. Quantification is shown in C-D. C-D. Percentage of cloNAT-resistant transformants displaying a mutant phenotype (pink cells, ade6; uracil auxotrophy and FOA resistance, ura4) after asynchronous (C) or G1-synchronized (D) wild-type cells were transformed with a sgRNA-loaded SpEDIT/pLSB (developed here, Figure 2) or pMZ379 (Rodríguez-López et al., 2016) plasmid targeting ade6+ or ura4+ (or no sgRNA plasmid as control). An HR template targeting the same or a different gene was cotransformed as indicated. n = number of cloNAT-resistant colonies assayed. Note that when an HR template targeting a different gene or no HR template was co-transformed into asynchronous cells, the number of cloNAT-resistant colonies obtained was drastically reduced. Experiment was repeated twice with similar results. E. For each condition in C-D, five colonies displaying the mutant phenotype (or 5 cloNAT-resistant colonies for no sgRNA plasmids) were taken and the gene targeted by the sgRNA was sequenced to confirm changes in its DNA sequence. Both ade6 and ura4 were sequenced when no sgRNA was used. Edited clones harbour the change contained in the corresponding HR template. Other mutations at PAM disrupt the PAM (NGG) sequence and the corresponding gene coding sequence. For asynchronous cells transformed with pLSB-ura4 (no HR template) and pMZ379-ade6 (no HR template) only two and one colonies were respectively obtained and analysed. * No colonies were obtained for these conditions. sgRNA, single guide RNA. PAM, proto-spacer adjacent motif. HR, homologous recombination donor template. N/S, non-selective medium. FOA, 5-fluoroorotic acid.
HR template ( Figure 3E, left). However, a variety of mutations that disable the PAM sequence and ultimately disrupt the coding sequence of each gene were detected in mutants generated when a non-homologous HR template (targeting a different gene to the sgRNA) or no HR template was provided ( Figure 3E, left).
A previous study suggested that G1 synchronization of S. pombe cultures by nitrogen starvation prior to CRISPR/ Cas9-mediated genome editing enhances transformation and deletion efficiencies (Rodríguez-López et al., 2016). The rationale for this was that in G1 only one copy of a target locus would need to be modified as opposed to the two copies that are present in G2 cells. The remodelled transcriptional programme of G1 cells was also expected to render genomic regions more open (Mata et al., 2002), and thus increase accessibility to the editing machinery.
Moreover, sequence analysis revealed that even when a matching HR template was co-transformed into G1 cells, not all mutant clones harboured the anticipated mutation that was presented by the HR template ( Figure 3E, right). This lack of accuracy is likely due to the suppression of the HR pathway that is known to occur in G1 cells ( Taken together, our results show that the SpEDIT system can generate targeted mutations at ade6 + and ura4 + with 100% mutagenesis efficiency using asynchronous cell cultures. Notably, our experiments show that G1 synchronization of S. pombe cells prior transformation has a detrimental effect on mutagenesis efficiency regardless of the CRISPR/Cas9 system used. SpEDIT shows reduced toxicity compared with the current pMZ379 CRISPR/Cas9 system in asynchronous cells To determine if the SpEDIT system leads to reduced toxicity compared to the current pMZ379 system, we measured colony area on selection plates after transforming asynchronous or G1-synchronized cultures with pLSB or pMZ379 plasmids targeting ade6 + or ura4 + (or empty plasmid controls) in the presence of a matching HR template. The colony area (equivalent to colony size) was found to be greater when asynchronous cells were transformed with pLSB as opposed to pMZ379 ( Figure 4A). The resulting difference in colony size was independent of the presence of a sgRNA, indicating that excessive levels of Cas9 alone, and not Cas9 targeting to a genomic locus, are sufficient to cause the observed toxicity ( Figure 4A). Consistent with this, the toxicity of adh1-Cas9 has also been shown to be independent of Cas9 catalytic activity (Ciccaglione, 2015).
In contrast, colony area measurements of pLSB and pMZ379 transformants obtained from G1-synchronized cells revealed no major difference in resulting colony size ( Figure 4B). This indicates that the toxicity related to high levels of catalytically active Cas9 is more prominent when transforming asynchronous cells. The lack of apparent toxicity in G1 cells is likely due to the known upregulation of non-homologous end joining-mediated repair and the suppression of homologous recombination repair that is known to occur at this cell cycle stage (Orthwein et al., 2015).  Figure 3). Colony area was quantified (cm 2 ) using ImageJ. sgRNA, single guide RNA. HR, homologous recombination donor template.
SpEDIT allows simultaneous editing at two nonhomologous genes at distinct locations in the S. pombe genome The availability of pLSB versions bearing different dominant selectable markers presented the opportunity to test if the simultaneous editing of two different non-homologous loci by co-transformation with and selection of two distinct pLSB plasmids expressing different sgRNAs was possible.
Sequencing of clr5 and meu27 in ten resulting cloNAT and hygromycin doubly resistant co-transformants revealed that two harboured both of the expected DNA changes. Five clones carried mutations in only one of the two targeted genes or bore mutations that uniquely affected the PAM sequence, and three clones displayed neither of the anticipated changes ( Figure 5C). Importantly, whole-genome sequencing of one of the isolates that contained both expected gene editing events revealed no additional genetic changes (SNPs or indels) in coding regions of the genome (Torres-Garcia et al., 2020).
These results demonstrate that our improved system can be utilised to perform simultaneous gene editing at two distant, non-homologous S. pombe loci, albeit with reduced efficiency relative to the frequency of editing of a single locus.
Seamless insertion, deletion and tagging at S. pombe loci using SpEDIT To assess the capabilities of SpEDIT in additional gene editing tasks, we utilised it to perform insertion, deletion and tagging at single S. pombe loci. Specifically, using SpEDIT, we inserted tetO binding sites downstream of the cup1 + (SPBC17G9.13c) gene ( Figure 6A). tetO binding sites allow tethering of proteins such as TetR-Clr4* and heterochromatin formation in the vicinity of the tethering site (Audergon et al., 2015;Ragunathan et al., 2015). A fusion-PCR construct containing 4xtetO sites with 120 bp homology arms flanking the desired insertion site was used as the HR template (see Table 1 for sgRNA and HR template sequences). Correct insertion of the cup1:4xtetO HR template resulted in ablation of the PAM sequence. Furthermore, we used SpEDIT to seamlessly fuse GFP in frame with the 3' end of the of cup1 + gene to produce Cup1-GFP ( Figure 6B). To generate the cup1-gfp HR template, the GFP open reading frame was amplified with oligonucleotides that had long extensions homologous to sequence immediately up-stream and downstream of the normal cup1 + STOP codon. This HR template also carried a DNA change (in the 5' long oligo) designed to disable the PAM sequence without altering the Cup1 protein sequence. Resulting strains were confirmed to carry the planned cup1:4xtetO or Cup1-GFP insertions in the absence of any associated selectable marker and have been utilised to show that heterochromatin-mediated silencing of cup1 + is sufficient to drive caffeine resistance in wild-type cells and that Cup1-GFP localises to mitochondria (Torres-Garcia et al., 2020). A. Schematic of experiment to simultaneously generate targeted point mutations in clr5 + and meu27+. Two sgRNA-loaded pLSB plasmids (with different selection markers) were co-transformed with two HR templates that create the desired point mutations and disable the corresponding PAM (NGG) sequence. Transformed cells were then selected on selective plates containing both cloNAT and hygromycin. B. sgRNA and HR template sequences for clr5 (left) and meu27 (right) are shown, along with Sanger sequencing chromatograms for a successfully edited clone. Full HR template sequences can be found in Table 1 Figure 6C). The epe1Δ HR template employed contained 80 bp arms homologous to sequences immediately flanking the epe1 + coding sequence as previously used for the deletion of other sequences (Rodríguez-López et al., 2016). Correct deletion of the epe1 + coding sequence results in loss of the sgRNA target and PAM sequences. In addition, we seamlessly inserted a sequence to encode a 3xFLAG epitope tag between the epe1 + gene promoter and the 5' end of the epe1 + coding sequence to allow the production of N-terminally 3xFLAG-tagged Epe1 without any associated selectable marker ( Figure 6D). To accomplish this, an in-frame epe1-3xflag-epe1 HR template containing 50 bp arms homologous to the sequence immediately flanking the epe1 + start codon was used. Correct insertion of the epe1-3xflag-epe1 sequence resulted in loss of both the sgRNA target and the PAM sequence. The resulting epe1Δ and 3xFLAG-Epe1 strains were recently utilised to study the role of Epe1 in ectopic heterochromatin island formation following caffeine exposure (Torres-Garcia et al., 2020). Notably, whole-genome sequencing of the cup1:4xtetO and epe1Δ strains revealed no additional genetic changes (SNPs or indels) in coding regions of the genome (Torres-Garcia et al., 2020).
In the four distinct genome editing scenarios described above, a maximum of eight primary transformants needed to be screened to obtain at least one that exhibited the desired Figure 6. SpEDIT allows seamless insertion, deletion and tagging at S. pombe loci. For sgRNA and HR template sequences see Table 1. A. 4xtetO binding sites were inserted downstream of cup1 + . Sanger sequencing chromatograms covering the insert junctions are shown for a successfully edited clone. PCR primers (half arrows) flanking the insert were used to amplify products from wild-type (wt) and edited strains. B. Cup1 was C-terminally tagged with a green fluorescent protein (GFP). Sanger sequencing chromatogram covering the gene-tag junction is shown for a successfully edited clone. Western blot using anti-GFP antibody was performed on wild-type (wt) and edited strains. C. The coding sequence of epe1+ was deleted. Sanger sequencing chromatogram covering the deletion junction is shown for a successfully edited clone. PCR primers (half arrows) flanking the deletion (and within the epe1+ coding sequence as control) were used to amplify products from wild-type (wt) and edited strains. D. Epe1 was N-terminally tagged with three FLAG epitopes. Sanger sequencing chromatogram covering the gene-tag junction is shown for a successfully edited clone. Western blot using anti-FLAG antibody was performed on wild-type (wt) and edited strains. sequence change. We therefore conclude that SpEDIT markedly speeds up the process of generating accurate insertion, deletion and tagging events at a variety of S. pombe loci.

Discussion
Here we report the development of SpEDIT, an optimized CRISPR/Cas9 editing system and method for the fission yeast, S. pombe (Figure 1 and Figure 2). SpEDIT makes use of Cas9 codon-optimised for expression in S. pombe that, coupled with the incorporation of a tDNA Ser /HDV ribozyme sgRNA expression cassette (Ryan et al., 2014), achieves 100% efficiency in generating mutations at targeted ade6 + or ura4 + genes in asynchronous cells (Figure 3). A high mutagenesis efficiency was also obtained with the pre-existing pMZ379 system in asynchronous cells (Jacobs et al., 2014;Rodríguez-López et al., 2016). However, SpEDIT displayed reduced toxicity by removing the detrimental physiological effects associated with high humanised Cas9 endonuclease expression and consequently speeds up the genome editing process (Figure 4). In addition, our analysis indicates that the use of G1-synchronized cell cultures for CRISPR/Cas9-mediated genome manipulation reduces the efficiency of targeted mutagenesis, with both the SpEDIT and pMZ379 systems, relative to asynchronous cultures (Figure 3). G1 synchronization therefore represents an unnecessary time-consuming step in the genome editing process.
SpEDIT can be used to introduce simultaneous mutations at two non-homologous genes at distinct locations in the S. pombe genome (Figure 5), and allows flexible engineering of seamless insertion, deletion and tagging events at S. pombe loci in the absence of linked selectable markers and without observed off-target sequence changes ( Figure 6). It is worth noting that many traditional S. pombe transformation protocols involve the use of carrier DNA. We advise against the use of carrier DNA as it has been shown to insert at many locations in resulting transformants, causing unplanned off-target mutations (Longmuir et al., 2019).
Besides achieving high mutagenesis efficiency, the greatest advance of SpEDIT is a very simple cloning protocol allowing sgRNA target sequences to be inserted with minimal effort Recently it was reported that homology arms of as short as 25 bp flanking each side of a cleavage site can be used to successfully introduce point mutations and epitope tags at S. pombe loci (Hayashi & Tanaka, 2019). Further analyses will be required to determine whether HR templates with such short homology arms are as efficient as longer arms when combined with SpEDIT. In addition, the tDNA/HDV ribozyme sgRNA expression cassette that was originally developed for S. cerevisiae has been used to express up to three tandem HDV-sgRNAs from a single tDNA RNAPIII promoter with 80% mutagenesis efficiency (Ryan et al., 2014). This suggests that a similar approach could be used with SpEDIT to simultaneously express multiple different sgRNAs that target a single locus or many distinct loci.
In summary, the combination of the CRISPR4P algorithm (Rodríguez-López et al., 2016), that conveniently aids the identification of suitable sgRNAs across the S. pombe genome, with SpEDIT, which provides a straightforward and user-friendly experimental method, markedly enhances the capabilities of CRISPR/Cas9-mediated genome editing in S. pombe. We anticipate SpEDIT will permit the broad application of genome editing procedures to fission yeast in order to explore diverse biological questions in this model fungal system.  Table 2.

Yeast strains and manipulations
Standard methods were used for fission yeast growth, genetics and manipulation (Moreno et al., 1991). S. pombe strains generated in this study are described in Table 3 To assess colony area of pLSB-or pMZ379-harbouring cloNAT-resistant colonies, plates were scanned after four days   Table 3. Schizosaccharomyces pombe strains used in this study.

Strain number Name Description
of incubation. Images were then analysed using ImageJ (v1.51) (analyse particles) with default settings.
Assessing mutations at the ade6 and ura4 loci Colonies harbouring mutations at the target genes ade6 + or ura4 + were identified through a replica-plating assay. cloNAT-resistant colonies were individually picked from YES plus cloNAT plates, re-streaked onto YES plates without selection and incubated at 32°C for two days. Isolates were then replicaplated onto the following plates: YES, YES 1/10 adenine (to examine ade6 + mutations), PMG minus uracil and PMG plus 5-fluoroorotic acid (to examine ura4 + mutations). Plates were incubated at 32°C for 2-4 days and then visually examined.
SpEDIT protocol A convenient protocol card of this procedure can be found by visiting allshirelab.com/spedit or by scanning the QR code in Figure 1.

Before you begin
• Download required DNA sequences at allshirelab.com/spedit, on Zenodo (Torres-Garcia, 2020) or by scanning the QR code in Figure 1 Required reagents • pLSB vector (75 ng/μL) -Available on request • NEB Golden Gate Assembly Kit (BsaI-HF v2) -NEB #E1601S • sgRNA fragment for Golden Gate assembly (1 ng/μL) -See below for design and preparation This project contains the following underlying data:   3 . For the expression of the sgRNA, they adopted an expression system successfully used in S. cerevisiae (Ryan et al. 2014) 4 , consisting of an RNA polymerase III-driven tRNA gene/hepatitis delta virus ribozyme (HDV) sequence that creates a defined 5' transcript and improves sgRNA stability and genome editing. The tDNA/HDV cassette further allowed insertion of two BsaI restriction sites for one-step cloning of the sgRNA by the Golding Gate cloning system. In addition, they inserted a GFP sequence as a placeholder within the BsaI sites for visual inspection of positive E. coli clones. Together, this modular design significantly improved the cloning procedure compared to previous methods that are hampered by the use of an inconvenient restriction enzyme (CspCI) or tedious PCR-based cloning methods. Moreover, by offering multiple expression vectors with different selection markers (natMX6, kanMX, or hphMX), SpEDIT allows versatile use in different strain backgrounds.
selection markers. However, the efficiency of these more challenging genetic alteration was significantly lower (about 10-20%) compared to introducing single point mutations.

Conclusion:
The CRISPR/Cas9 system presented here overcomes previous technical hurdles and significantly improves the usability of Cas9 based genome editing in S. pombe. The manuscript is well written, and the protocol is easy to follow. The authors have shared their system with my lab, and we have successfully used it to introduce mutations into several genes, confirming the efficiency of this system. Though, we did notice that editing at some loci is more challenging. Previous studies (e.g., Rodriguez-Lopez et al., 2016) 2 have analyzed a larger number of genomic targets and compared different sgRNAs to assess the efficiency of the editing method. Thus, the reported high efficiency for ade6+ and ura4+ may not always be achieved for other genomic targets. Furthermore, while several parameters have been changed compared to existing methods, it is sometimes unclear how these changes (e.g., codon-optimization, reduced expression) have improved efficiency (see comments below). Maybe, the authors could provide more information in a revised version, which may clarify some of the points mentioned below.

Comments and questions:
For comparing SpEDIT with a previous method reported by the Bähler Lab (Rodriguez-Lopez et al., 2016) 2 , the authors assessed mutagenesis efficiency and correct editing in Figure 3C, D. When using matching sgRNAs and HR templates, both methods seem to work equally efficiently for asynchronous cells. In contrast, differences were found for mismatching or no templates, as demonstrated by plotting mutant phenotypes as percentage relative to the total number tested. However, we wonder whether plotting the actual number of clones with mutant phenotypes would be more informative than plotting percentage. While percentage includes information on false positives, which is in principle meaningful, this analysis may be skewed when analyzing different total numbers of samples, especially for low numbers. For instance, for analyzing ade6 mutations, both pLSB with pMZ379 resulted in similar numbers of clones with mutant phenotypes when using the correct HR template (150 vs. 150), non-correct HR template (12 vs. 26), or no HR template (0 vs. 1). However, by plotting percentages (i.e., number of clones with mutant phenotypes relative to total) the difference between both plasmids seems much larger (e.g., 0% vs. 100% for no template). This may reflect noise due to low sample numbers, but this does not immediately become clear when looking at the graphs. The same applies to the tables shown in Figure 3E. Also, providing here exact numbers (e.g., 5/5 or 4/5) seems more accurate and informative (especially when only 1 or 2 clones have been tested).

○
The authors emphasize that they codon-optimize the Cas9 sequence and that its reduced expression results in decreased toxicity; however, how does this contribute to improved editing or a more user-friendly procedure? While the pLSB plasmid decreased toxicity in asynchronous cells compared to pMZ379, this did not improve gene editing efficiency for inserting single point mutations in ura4+ and ade6+. Did the authors also compare the two plasmids for editing other targets or introducing more rate-limiting changes (e.g., deletions or insertions)?
○ What is the expression strength of adh15 promoter? (for comparison: for the adh1 gene, 400 RNA molecules/cell have been reported).
○ How does using electroporation affects transformation efficiency compared to chemical transformation? Is there a difference in the number of colonies after transformation with pLSB vs. pMZ379? Along the same line, to what extent is the comparison of gene editing using asynchronous and G1 synchronized cells confounded by using different transformation methods? ○ © 2020 Martin S. This is an open access peer review report distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Sophie Martin
Universite de Lausanne, Lausanne, Switzerland This manuscript by Torres-Garcia and colleagues presents an improved CRISPR/Cas9 method for genome engineering in the fission yeast S. pombe. The main improvements over previously described methods are reduced toxicity of the Cas9 plasmid and greater ease of cloning of the sgRNA. They show that the method is highly efficient for introduction of point mutations and can also be used to introduce point mutations in two loci simultaneously or to insert or delete longer sequences.
The manuscript is well presented and convincing. The authors have already distributed their reagents, and my lab has used them with success. I can thus confirm that the method is functional and easy to use. It is likely to become widely used in the community. My comments below are mainly asking for some clarification.
One question that is not addressed is why Cas9 is toxic. The readout is small colony size, but whether this is linked to Cas9 binding the genome and/or inducing cuts is not clear. The authors cite a master thesis, which reports that toxicity is not linked to Cas9 catalytic activity, which is interesting, but to my knowledge this has not been probed further. I would suggest spelling out that the reason for toxicity is unclear and incite researchers to use the method with care.
I was also puzzled by why the target site appears to be more efficiently mutated when cells are cotransformed with an irrelevant HR template than with no HR template ( Figures 3C-D). Any thought about this?
What is the efficiency of sequence insertion/tagging/deletion? The text states that 8 clones needed to be analysed to find at least 1 positive one. This suggests lower success rate than for the point mutations. It would be informative to state the results fully and make this point clear.

Regarding the methodology:
The authors used Golden Gate to clone the sgRNA sequence in pLSB, but in principle other cloning strategies (including standard T4-ligase-based strategies) are possible. This would be good to indicate. ○ A similar point goes for the transformation strategy. The authors state in the text that they used electroporation, which is not the most widely used strategy in yeast labs. However, in the protocol, they state that any transformation strategy can be used. Does this matter? ○ In Figure 2C, I would suggest modifying the scheme to better indicate the difference of colour of the colonies (rather than cultures). The scheme seems to indicate that all will be picked and grown, and that the green cultures will be discarded, but in reality, one only picks from the non-green ones.

Regarding data accessibility:
Reviewer Expertise: Fission yeast molecular genetics.
I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.